Postdoctoral Researcher - Deep Learning for Computer Vision

About the Role

INRIA Grenoble Rhone-Alpes is recruiting a postdoctoral researcher to work within the Thoth research team on deep learning methods for visual scene understanding, 3D reconstruction, and video analysis. The project is funded through a Horizon Europe collaborative grant. Contracts are for 18 to 24 months.

Salary follows the INRIA postdoctoral pay scale: EUR 2,788 to 3,600 per month gross, depending on years of experience after the PhD, plus meal vouchers and full French social security and pension coverage. INRIA Grenoble is one of the leading AI and computer vision research sites in Europe, co-located with Université Grenoble Alpes.

Responsibilities

• Design and implement novel deep learning architectures for one or more of: dense 3D reconstruction from video, semantic scene segmentation, or self-supervised visual representation learning
• Run ablation experiments on standard benchmarks (ScanNet, Waymo, KITTI, or similar) with rigorous reproducibility reporting
• Publish results at top venues: CVPR, ICCV, ECCV, NeurIPS, or ICLR
• Contribute to EU project deliverables and attend consortium meetings (one to two trips per year within Europe)
• Co-supervise PhD students in the Thoth team

Requirements

• PhD in computer science, electrical engineering, or applied mathematics, with a focus on computer vision or machine learning
• At least two first-author publications at CVPR, ICCV, ECCV, NeurIPS, ICML, or ICLR
• Proficiency in Python and PyTorch (or equivalent deep learning framework)
• Good written and spoken English; French is not required but is useful for daily life in Grenoble

Nice to Have

• Experience with 3D representations (NeRF, Gaussian splatting, point clouds)
• Background in video understanding or vision-language models
• Familiarity with EU Horizon project administration

How to Apply

Apply online at jobs.inria.fr by 15 July 2026. Upload your CV, a research statement (max 2 pages outlining your planned contribution to the team’s agenda), a full publication list with links, and names and email addresses of two academic referees. The Thoth team will acknowledge receipt within one week and schedule a video interview with shortlisted candidates.